Sign in Sign up

Learning path

Deep Learning Foundations

Neurons and layers, your first Keras network, CNNs for images, RNNs and LSTMs, and PyTorch, with animated films and runnable Python. Part of the Data Scientist premium track.

Start with lesson 1
  1. 1

    Neurons, Layers and How a Neural Network Learns

    Build a neural network from scratch in NumPy: neurons, weights, activation functions, forward pass, loss, and backpropagation, and watch it learn XOR, a problem a straight line can't…

    3 min read
  2. 2

    Your First Neural Network with Keras (TensorFlow)

    Build, train and evaluate a neural network with Keras: Sequential models, Dense layers, compile, fit with validation, callbacks like EarlyStopping, and a runnable scikit-learn MLP version.

    3 min read
  3. 3

    CNNs for Images: Convolution, Pooling and Transfer Learning

    How convolutional neural networks see images: filters that detect edges, pooling, feature maps, a Keras CNN, and transfer learning with pretrained models like ResNet. Convolution runs in NumPy.

    3 min read
  4. 4

    RNNs and LSTMs for Sequences

    Networks with memory: how recurrent neural networks read sequences step by step, why plain RNNs forget, how LSTM and GRU gates fix it, and an LSTM forecaster in…

    3 min read
  5. 5

    The Same Models in PyTorch

    Learn PyTorch, the framework behind most AI research and LLMs: tensors, autograd, nn.Module, the training loop, DataLoaders and moving to a GPU, compared side by side with Keras.

    3 min read